IO.NET was exclusively devoted to developing institutional-grade quantitative trading systems for both the United States stock market and the cryptocurrency markets.

The steep cost of GPU cards posed a major challenge. Requiring over 50 cards monthly, our expenses surpassed $100K, And for us like most other ML startups, this created a substantial financial burden.

This infrastructure needs a team of MLOps and DevOps. Ray's discovery cut our backend development time from six months to 60 days. When we integrated Ray and deployed our GPU clusters, high GPU costs from on-demand cloud providers became a major obstacle.

Confronted with traditional computing limits, we explored Distributed Computing and created a unique solution. This led to decentralized clusters within a DePIN (Decentralized Physical Infrastructure Network).